Assessing Research on Student Evaluations

This is a really useful article.

I found this point particularly convincing:

Are ratings affected by situational variables?

The research says that ratings are robust and not greatly affected by situational variables. But we must keep in mind that generalizations are not absolute statements. There will always be some variations. For example, we know that required, large-enrollment, out-of-major courses in the physical sciences get lower average ratings than elective, upper-level, major courses in virtually all other disciplines. Does this mean that teaching quality varies? Not necessarily. What it does show is that effective teaching and learning may be harder to achieve under certain sets of conditions. There is a critical principle for evaluation practice embedded here: to be fair, comparisons of faculty teaching performance based on ratings should use sufficient amounts of data from similar situations. It would be grossly unfair to compare the ratings of an experienced professor teaching a graduate seminar of ten students to the one-time ratings of a new instructor teaching an entry-level, required course with an enrollment of 300.

To me, this suggests the obvious: That large lectures are not as successful as small classes.

Also obviously, this comes down to a funding issue.

Nevertheless.